Skip to content

Remove duplicate code for processing arrays vs strings #79233

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 17 commits into from
Jul 9, 2025
Merged

Remove duplicate code for processing arrays vs strings #79233

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
9256667
Use spans in low level lexer char array handling code
CyrusNajmabadi Jul 2, 2025
900c83a
Use spans in low level lexer char array handling code
CyrusNajmabadi Jul 2, 2025
7b4f2e6
VB tests
CyrusNajmabadi Jul 2, 2025
04f25de
Remove duplicate code for processing arrays vs strings
CyrusNajmabadi Jul 2, 2025
0bc7247
Revert
CyrusNajmabadi Jul 2, 2025
928490f
Fix test
CyrusNajmabadi Jul 2, 2025
7ee4ca0
Merge branch 'useSpan' into shareStringSpanCode
CyrusNajmabadi Jul 2, 2025
b43655e
Share more code
CyrusNajmabadi Jul 2, 2025
ef4563a
Remove unused function
CyrusNajmabadi Jul 2, 2025
a271448
Merge branch 'main' into shareStringSpanCode
CyrusNajmabadi Jul 2, 2025
ad4e5c5
Update src/Compilers/Core/Portable/InternalUtilities/StringTable.cs
CyrusNajmabadi Jul 2, 2025
054b38c
Remove using
CyrusNajmabadi Jul 2, 2025
749ed47
update
CyrusNajmabadi Jul 2, 2025
2ef177d
Merge branch 'main' into shareStringSpanCode
CyrusNajmabadi Jul 3, 2025
15d2c56
Simplify
CyrusNajmabadi Jul 5, 2025
16e0c6e
Merge branch 'main' into shareStringSpanCode
CyrusNajmabadi Jul 7, 2025
ccdbc08
Use collection expression
CyrusNajmabadi Jul 7, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 2 additions & 11 deletions src/Compilers/Core/Portable/InternalUtilities/Hash.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -286,9 +286,7 @@ internal static int GetFNVHashCode(string text, int start, int length)
=> GetFNVHashCode(text.AsSpan(start, length));

internal static int GetCaseInsensitiveFNVHashCode(string text)
{
return GetCaseInsensitiveFNVHashCode(text.AsSpan(0, text.Length));
}
=> GetCaseInsensitiveFNVHashCode(text.AsSpan());

internal static int GetCaseInsensitiveFNVHashCode(ReadOnlySpan<char> data)
{
Expand Down Expand Up @@ -375,14 +373,7 @@ internal static int GetFNVHashCode(char ch)
/// <param name="text">The string to combine</param>
/// <returns>The result of combining <paramref name="hashCode"/> with <paramref name="text"/> using the FNV-1a algorithm</returns>
internal static int CombineFNVHash(int hashCode, string text)
{
foreach (char ch in text)
{
hashCode = unchecked((hashCode ^ ch) * Hash.FnvPrime);
}

return hashCode;
}
=> CombineFNVHash(hashCode, text.AsSpan());

/// <summary>
/// Combine a char with an existing FNV-1a hash code
Expand Down
216 changes: 11 additions & 205 deletions src/Compilers/Core/Portable/InternalUtilities/StringTable.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,12 +5,8 @@
using System;
using System.Text;
using System.Threading;
using Microsoft.CodeAnalysis.PooledObjects;
using Microsoft.CodeAnalysis.Collections;

#if DEBUG
using System.Diagnostics;
#endif
using Microsoft.CodeAnalysis.PooledObjects;

namespace Roslyn.Utilities
{
Expand Down Expand Up @@ -158,72 +154,10 @@ internal string Add(ReadOnlySpan<char> chars)
}

internal string Add(string chars, int start, int len)
{
var hashCode = Hash.GetFNVHashCode(chars, start, len);

// capture array to avoid extra range checks
var arr = _localTable;
var idx = LocalIdxFromHash(hashCode);

var text = arr[idx].Text;

if (text != null && arr[idx].HashCode == hashCode)
{
var result = arr[idx].Text;
if (StringTable.TextEquals(result, chars, start, len))
{
return result;
}
}

string? shared = FindSharedEntry(chars, start, len, hashCode);
if (shared != null)
{
// PERF: the following code does element-wise assignment of a struct
// because current JIT produces better code compared to
// arr[idx] = new Entry(...)
arr[idx].HashCode = hashCode;
arr[idx].Text = shared;

return shared;
}

return AddItem(chars, start, len, hashCode);
}
=> Add(chars.AsSpan(start, len));

internal string Add(char chars)
{
var hashCode = Hash.GetFNVHashCode(chars);

// capture array to avoid extra range checks
var arr = _localTable;
var idx = LocalIdxFromHash(hashCode);

var text = arr[idx].Text;

if (text != null)
{
var result = arr[idx].Text;
if (text.Length == 1 && text[0] == chars)
{
return result;
}
}

string? shared = FindSharedEntry(chars, hashCode);
if (shared != null)
{
// PERF: the following code does element-wise assignment of a struct
// because current JIT produces better code compared to
// arr[idx] = new Entry(...)
arr[idx].HashCode = hashCode;
arr[idx].Text = shared;

return shared;
}

return AddItem(chars, hashCode);
}
=> Add([chars]);

internal string Add(StringBuilder chars)
{
Expand Down Expand Up @@ -260,39 +194,7 @@ internal string Add(StringBuilder chars)
}

internal string Add(string chars)
{
var hashCode = Hash.GetFNVHashCode(chars);

// capture array to avoid extra range checks
var arr = _localTable;
var idx = LocalIdxFromHash(hashCode);

var text = arr[idx].Text;

if (text != null && arr[idx].HashCode == hashCode)
{
var result = arr[idx].Text;
if (result == chars)
{
return result;
}
}

string? shared = FindSharedEntry(chars, hashCode);
if (shared != null)
{
// PERF: the following code does element-wise assignment of a struct
// because current JIT produces better code compared to
// arr[idx] = new Entry(...)
arr[idx].HashCode = hashCode;
arr[idx].Text = shared;

return shared;
}

AddCore(chars, hashCode);
return chars;
}
=> Add(chars.AsSpan());

private static string? FindSharedEntry(ReadOnlySpan<char> chars, int hashCode)
{
Expand Down Expand Up @@ -330,39 +232,7 @@ internal string Add(string chars)
}

private static string? FindSharedEntry(string chars, int start, int len, int hashCode)
{
var arr = s_sharedTable;
int idx = SharedIdxFromHash(hashCode);

string? e = null;
// we use quadratic probing here
// bucket positions are (n^2 + n)/2 relative to the masked hashcode
for (int i = 1; i < SharedBucketSize + 1; i++)
{
e = arr[idx].Text;
int hash = arr[idx].HashCode;

if (e != null)
{
if (hash == hashCode && TextEquals(e, chars, start, len))
{
break;
}

// this is not e we are looking for
e = null;
}
else
{
// once we see unfilled entry, the rest of the bucket will be empty
break;
}

idx = (idx + i) & SharedSizeMask;
}

return e;
}
=> FindSharedEntry(chars.AsSpan(start, len), hashCode);

private static string? FindSharedEntryASCII(int hashCode, ReadOnlySpan<byte> asciiChars)
{
Expand Down Expand Up @@ -400,38 +270,7 @@ internal string Add(string chars)
}

private static string? FindSharedEntry(char chars, int hashCode)
{
var arr = s_sharedTable;
int idx = SharedIdxFromHash(hashCode);

string? e = null;
// we use quadratic probing here
// bucket positions are (n^2 + n)/2 relative to the masked hashcode
for (int i = 1; i < SharedBucketSize + 1; i++)
{
e = arr[idx].Text;

if (e != null)
{
if (e.Length == 1 && e[0] == chars)
{
break;
}

// this is not e we are looking for
e = null;
}
else
{
// once we see unfilled entry, the rest of the bucket will be empty
break;
}

idx = (idx + i) & SharedSizeMask;
}

return e;
}
=> FindSharedEntry([chars], hashCode);

private static string? FindSharedEntry(StringBuilder chars, int hashCode)
{
Expand Down Expand Up @@ -469,39 +308,7 @@ internal string Add(string chars)
}

private static string? FindSharedEntry(string chars, int hashCode)
{
var arr = s_sharedTable;
int idx = SharedIdxFromHash(hashCode);

string? e = null;
// we use quadratic probing here
// bucket positions are (n^2 + n)/2 relative to the masked hashcode
for (int i = 1; i < SharedBucketSize + 1; i++)
{
e = arr[idx].Text;
int hash = arr[idx].HashCode;

if (e != null)
{
if (hash == hashCode && e == chars)
{
break;
}

// this is not e we are looking for
e = null;
}
else
{
// once we see unfilled entry, the rest of the bucket will be empty
break;
}

idx = (idx + i) & SharedSizeMask;
}

return e;
}
=> FindSharedEntry(chars.AsSpan(), hashCode);

private string AddItem(ReadOnlySpan<char> chars, int hashCode)
{
Expand All @@ -512,17 +319,16 @@ private string AddItem(ReadOnlySpan<char> chars, int hashCode)

private string AddItem(string chars, int start, int len, int hashCode)
{
// Don't defer to ReadOnlySpan<char> here, as it would cause an extra allocation
// in the case where start/len exactly match the full span of chars.

var text = chars.Substring(start, len);
AddCore(text, hashCode);
return text;
}

private string AddItem(char chars, int hashCode)
{
var text = new String(chars, 1);
AddCore(text, hashCode);
return text;
}
=> AddItem([chars], hashCode);

private string AddItem(StringBuilder chars, int hashCode)
{
Expand Down